Using the Webrecorder
The Electronic Literature Lab [ELL], the site where my research took place, partnered with the organization Rhizome to use its Webrecorder tool to preserve The Progressive Dinner Party. Developed by IIya Kreymer and directed by Dragan Espenshied, the project provides an archiving tool that preserves page performance and functionality by emulating the web browser. Members of the Electronic Literature Lab held a Zoom meeting with Dragan Espenshied remotely, who introduced them to the Webrecorder and answered any questions the lab may have had. One of Espenshied’s main points was that there are two methods of recording that are made available by the Webrecorder: the first is used for local files--the original files held locally on ELL's computer; the second is reserved for external files, or those works accessed through external links.
I obtained local files from the artists by contacting them via email, while Dene Grigar sought them out through contact with the artists via Facebook and Instagram. Though we had many of the artists’ email addresses already, I had to find some of their contacts by browsing the web for online portfolios, artist statements, and interviews, or asking assistance from Ley and Luesebrink. Once contact was made, the artist would search her personal archives for her works and send me the files, and I would proceed to download them to my computer and check for possible missing files. If files were missing, I would make a list of what was needed and send it to the artist, who would then continue searching for more files. On occasion, I would have to make alterations to the work to make them functional (such as reconstructing folders to match file directories found in the code.)
Once all the files were obtained, I would locally run them through the Webrecorder. However, certain steps had to be taken prior to the recording. The local files first had to be prepared by placing them all in the same directory, and the URLs had to be given newly generated links that differed from the orginal site's. The WARCIT utility was then used to convert the contents of the directory into the web archive format known as WARC (WebARChive.) Afterwards, the generated WARC file was compared with the original files to check for accuracy, and notes were placed in Scalar whenever discrepancies occurred. Lastly, the WARC file was published to the online Webrecorder and made accessible online. When necessary, the newly generated files were compared to the original ones loaded on a legacy computer in the lab.
When needed, I was able to find the external files to a work still intact on the Web. I oftentimes found links to these files in The Progressive Dinner Party, and could run them in The Wayback Machine if need be. If I could not obtain them this way, I would search the web for other possible locations the work could have been housed, including in online journals and portfolios. Once the external files were obtained, they were first examined to ensure their functionality. I would traverse the work and compare it to other references (such as images or video traversals) of what they looked like in their original state and environment. If a work appeared differently through Chrome, Safari, FireFox or Pale Moon browsers, I made a note in the website regarding the differences between the newly recorded work and the original. When I finished checking the work for its functionality, I would copy and paste the link into the online Webrecorder. A compatible browser was then chosen from the ones made available in the Webrecorder. Once these steps were complete, I manually traversed the works for all possible links on the site to capture them in full. The Webrecorder would generate a new URL so that the files were made available online.
In some instances, neither the local nor external files of a work could be obtained. This either meant that I could not make contact with the artist or that the files were no longer available to anyone.
Below is a chart by ELL manager Holly Slocum detailing the process in which works are captured by the Webrecorder.